Very cool! You could also possibly constrain the problem even further so that the hight of the paddle when hitting the ball is proportional to the note pitch. In that case, and if using a single instrument, the resulting video would be an encoding of the song.
I like it! We probably don't have enough degrees of freedom to assign absolute pitch to heights since large jumps in pitch will be unreachable. However encoding relative pitch would be simple, currently I pick the sign of the angle randomly. We could make it positive when ascending in pitch and negative when descending.
Very cool! You could also possibly constrain the problem even further so that the hight of the paddle when hitting the ball is proportional to the note pitch. In that case, and if using a single instrument, the resulting video would be an encoding of the song.
I like it! We probably don't have enough degrees of freedom to assign absolute pitch to heights since large jumps in pitch will be unreachable. However encoding relative pitch would be simple, currently I pick the sign of the angle randomly. We could make it positive when ascending in pitch and negative when descending.